rlm-cli

a CLI implementation of Recursive Language Models (arXiv:2512.24601). instead of stuffing your entire context into one LLM call, rlm writes code to actually process the data - slicing, chunking, running sub-queries on pieces and looping until it gets the answer. works with claude, gpt, gemini, auto-loads your project file tree as context. npm i -g rlm-cli and you’re running in 30 seconds. 190+ stars on GitHub.

fim-coder-model

a FIM (fill-in-the-middle) training framework with AST-aware extraction for code completion. extracts semantic boundaries from Rust source files, generates targeted training samples, and fine-tunes using QLoRA. the idea is that AST-aware extraction gives you better training samples than just splitting by lines.